Skip to main content

Datasets

Standard Dataset

Weibo and Twitter

Citation Author(s):
bingxin li (North China University of Technology)
Submitted by:
bingxin li
Last updated:
DOI:
10.21227/6cbb-cx13
Data Format:
666 views
Categories:
Keywords:
No Ratings Yet

Abstract

Weibo and Twitter

1)The Weibo dataset is derived from the Weibo social platform. The collection of true information in this dataset originates from authoritative Chinese sources, while fake information is acquired through the official Weibo rumor suppression system. Each data instance within this dataset comprises both a news text and a corresponding news image.

2)The Twitter dataset is sourced from the Twitter social platform. Each tweet's content consists of a concise text message and additional images or videos. The development set encompasses around 6,000 rumors and 5,000 non-rumor tweets, stemming from 11 rumor-related events. Each data instance in this dataset includes both a news text and a corresponding news image.

They are split into 7:1:2 ratios for training, validation, and test sets.

Instructions:

Weibo and Twitter